Journal of Education and Learning; Vol. 8, No. 5; 2019 ISSN 1927-5250 E-ISSN 1927-5269

Published by Canadian Center of Science and Education

31

POE: Understanding Innovative Learning Places and Their Impact on Student Academic Engagement—Index 6–8 ‘Alpha’ Survey

Developments Lennie Scott-Webber1, Roger Konyndyk2 & Marilyn Denison3

1 INSYNC: Education Research + Design, Estero, Florida, USA 2 Statistical Consulting, Grand Rapids, Michigan, USA 3 DLR Group, K12 Education Practice, Dallas, Texas, USA Correspondence: Lennie Scott-Webber, PhD, INSYNC: Education Research + Design, Estero, Florida, 33929, USA. E-mail: lenniesw.insync@yahoo.com Received: August 1, 2019 Accepted: August 28, 2019 Online Published: September 19, 2019 doi:10.5539/jel.v8n5p31 URL: https://doi.org/10.5539/jel.v8n5p31 Abstract New evidence builds upon the Student Engagement IndexTM and Teacher Engagement IndexTM research (Scott-Webber, Konyndyk, & French, 2019; Scott-Webber, Konyndyk, French, & French, 2018; Scott-Webber, Konyndyk, French, Lembke, & Kinney, 2017) determining post-occupancy answers to, “Can we demonstrate that the design of the built environment for grades 6–8 impacts student academic engagement levels post-occupancy?” The early studies used respondents from grades 9–12. This one is from users in grades 6–8 (‘alpha’ pilot). All studies were conducted in the USA as convenience samples. Engagement performance is a high predictor of student success across multiple domains and learning/work experiences. Specifically, “Research that shows that engagement, the time and energy students devote to educationally purposeful activities, is the best single predictor of their learning and personal development” (Anonymous, NSSE, 2010, p. 2), and thus our research focus. From both the students and educators perspectives, design of the built space impacts engagement performance (p < .0001). Keywords: academic engagement, architecture, learning place design, post-occupancy evaluation, survey development 1. Introduction 1.1 Does Design Make a Difference in Learning and Teaching? Why survey students and educators to see if design makes a difference in their everyday learning or teaching situations? Because, evidence indicates engagement performance is a high predictor of success across multiple domains and learning/work experiences. Specifically, “Research shows that engagement, the time and energy students devote to educationally purposeful activities, is the best single predictor of their learning and personal development” (Anonymous, NSSE, 2010, p. 2), and thus our research focus. Furthermore, and perhaps just as importantly, most of our human experiences are inside this built ‘box’ called school; the USA average is 6.5 hrs per day for approximately 180 days = 1,170 hrs per year; added up = 15,210 hours. Therefore, it stands to reason that where we spend our time and how these places are designed to support individual needs is critical to understand. This current work builds on a career effort and the questions used are framed from multiple researchers in several domains in a holistic approach titled the Users Environmental Interaction Framework.v2© (UEIF.v2) (Scott-Webber, 1999). This study builds upon that work in trying to understanding how the deign of the built environment impacts student academic engagement levels (Scott-Webber, Konyndyk, French, & French, 2018; Scott-Webber, Konyndyk, French, Lembke, & Kinney, 2017; Scott-Webber, 2004; Scott-Webber, Marini, & Abraham, Spring, 2000). This report differs as it studies a new age cohort - students in grades 6 to 8. The research question for this study continues to be the same as for the higher grades of 9 to 12. It is, “Can we demonstrate that the design of the built environment for grades 6-8 impacts student academic engagement levels post-occupancy?” The research design is explained next.

jel.ccsenet.o

This reseapost-occupimpacts huof the desqualitativeuses a mixand qualita

1) Qu2) Qunaturinqui23).

The researinquiry relthe contex1.2 ExplorAs we resenew, we sAlthough mis that “…they have Our researdesign of their respounderstandstudents, (supportinglearning sp

• Red

• P• A

asen

• C

org

arch team utipancy. The fouman behaviorsign solution. Me and quantitatxed-method (Jative) for compuantitative techualitative techre, and may iiry, cultural pr

rch question alating to how

xt of the educatring the Imporearch new dessurvey post-ocmany definitio

…POE can be dbeen built and

rch over multithe school ma

onses differ slid: (1) the leve(2) whether sg the educationpaces. This stu

Research-grounductional levelerceptions rec

Awareness of ts the individungagement

Comparison to

lized a Humacus for humanr, importantly Multiple methtive. To ensureJohnson, Onwuparative purpohniques = liter

hniques = “obsinclude particirobes, and oth

Figure 1. A h

and the researthe built envirtional setting ftance of Studyign concepts fccupancy at aons of Post-Ocdefined as ‘thed occupied foriple years has akes a significightly. Engageel of academicstudents at granal efforts, andudy followed anded questionls across multieived from bohe building’s

ual learning sp

instructional s

Journal of Ed

an-Centered Rn-centered reshow users beh

hods are often e a rich set of uegbuzie, & T

oses, and thus lrature review, serving and talipant observat

her methods d

human-centere

rch design are ronment impacfor grade levelsying the Problefor education, aa minimum ofccupancy Evalu process of evr some time’ (indicated that cant differencement is a critc engagement ades 6–8, a nd then (4) howa rigorous sciennnaire proven iple years th user groups(macro or ove

paces’ (micro o

strategies and h

ducation and Le

32

Research Desiearch understahave and if tha

used to answf data is generaTurner, 2007) rlimits research“questionnaireking to peopletion, artifact esigned to sam

ed research des

in harmony acts human behs 6 to 8 and usem from a Postand in order to

f at least threeuation (POE) h

valuating build(Preiser, Harvfrom both the

e relative to aical piece of tas perceived b

new cohort lew this knowledntific research

to be both

s, students and erall/rest of theor classrooms

how these buil

earning

ign (HCRD) ands how the at behavior wa

wer research quated, the HcRDresearch desigh bias. The HcRes and surveyse…these methanalysis, photmple human e

sign protocol (

as the HcRD havior. Our ressed post-occupt-Occupancy Po mitigate the e months afterhave been prop

dings in a systevey, & Edwarde students’ andacademic engathis research. Tby both studenvel can disce

dge may be useprotocol and preliable and

educators, aboe built environ) design and h

lt spaces suppo

protocol to stdesign of the

as anticipated iuestions, but fD (see figure n (inclusive oRD method us” (Hanington, ods are typicato and diary experience” (H

(HcRD)

is human-censearch questio

pancy. Perspective effect of experr occupation iposed, a usefuematic and rigd, 1988,” in, Ad the educatoragement perfoThis work hasnts and by edu

ern this impaced to impact dprovided the fovalid, pre-tes

out academic enment) buildinhow they cont

ort these endea

Vol. 8, No. 5;

tudy this que built environin the developfall into two a1) protocol al

of both quantitses the followin

2010, p. 22), aally ethnographstudies, conte

Hanington, 201

ntered and supon asks the sam

riencing somein a new buil

ul, classic definorous manner

Anonymous, 2rs’ perspectiveormance. How the opportuniucators about

ct, (3) if desigdesigns for theollowing: sted at a diff

engagement leng’s design, astribute to leve

avors.

2019

estion ment ment

areas, ways tative ng: and hic in xtual

10, p.

ports me in

thing ding.

nition after

008). es the

wever, ity to their

gn is next

ferent

vels well

els of

jel.ccsenet.o

Overall, thknowledge(overall/rework to bu1.3 RelevaThe Stude

builda hol1999)interathe dDimeZoneEnvir(Hall2001;

The curreStricklandfor gradesdesigned tenvironmeasks how wof choice environmebe comfor[the situatithe valuesquestions. The Teacheducator tperspectiveducator hdigs in deesupport theThis reseawith each

org

his 6–8 ‘alphe that the desiest of built learuild reliable anant Scholarshipnt Engagemens on a career eistic approach). A review action/engagemdesign of the ensions of Vals at the microronment Behav, 1966; Somm; Scott-Webbe

Fig

ent study build, & Kapitula, 2s 9–12 survey to ask a series ent) (2) the clawell one can mand control

ental qualities rtable and be cion of each scs placesd on

her Engagemeto rate his/her

ve is critical ahow the designeper in the eduem.

arch was attemother, with th

a’ pilot (3 scign and use ofrning spaces) and valid instrump nt IndexTM (SEeffort and the qh titled the Useof the framew

ment understanbuilt spaces—

lue and Enviroo level. This frvior Theorists

mer, 1959; Maer, 2000 & 200

gure 2. Users e

lt upon early 2013; Scott-Winstruments.

of questions reassroom (a micmove about, naover where aare touched upconnected to ochool, and its odifferent type

nt IndexTM (Tr student’s en

and the teachen of the built ucators perspec

mpting to ‘provheir teachers, a

Journal of Ed

chools, studenf the built spaareas impacts ments for whic

EI) questionnaiquestions useders Environmework is nextndings with mu—the micro lonment, (3) twframework has, in an effort t

aslow, 1943; B04; Scott-Webb

environmental

work done Webber, Marini

This newest related to the Ucro environmeavigate from inand how studpon; accessibliothers; and moorganization’ses of learning

TEI) questionaingagement lever responses aspace is suppoctive to unpack

ve’ that the deand with the ac

ducation and Le

33

nts n = 2,007ace at both thestudent academch to study this

ire is focused od are framed frental Interactio. The graphicultiple facets ievel, or class

wo Responses s built on the rto more fully Bloom, Krathwber, Abraham,

l interaction fra

in higher edui, & Abraham,research effort

UEIF.v2 relativent), or learninnstructional strents may wisity to ‘tools’ wotivational facs culture] of thg experiences.

ire follows thevels using theare compared orting his/her k the instructio

sign of space cademic conte

earning

; teachers n =e micro (classrmic engagemes important qu

on the studentfrom multiple ron Frameworkc in Figure 2including thesesroom, and mof Internal an

research of mexamine spacewohol, & Har& Marini, 200

amework.v2©

ucation (Scott Spring, 2000)t studies gradeve to two spating spacebuilt erategy to instrush to learn anwithin their leactors. It then fhe school and

Finally, it en

e SEI with twe same questito the studentneeds, again a

onal strategies

makes a diffeent. This work

= 210) continroom/learning ent levels. Thisuestion.

’s perspective.researchers in k.v2© (UEIF.v2 represents te specific segm

macro level, ond Behavioral,

many others, pae and its relatirrow, 1956; E00) (see Figure

© (UEIF.v2)

t-Webber, 20), and the relates 6 through 8ial areas (1) thenvironment ofuctional strategnd with whomarning spaces; focuses on thetheir perceivends with a se

wo specific focions the studet responses. Sat the macro aand how the d

erence in how tried to under

Vol. 8, No. 5;

nues to extendplace), and m

s team continu

. This current several domai

v2) (Scott-Wethe complexitments: (1) layeor overall, (2), and (4) Proxarticularly clasionship to its lliot & Covine 2).

14; Scott-Wetively new rese8. The surveye overall (or mf the school. Itgy; the opportm; whether inability to see, situational cud understandinet of demogra

ci. First, it askents’ receive. Second, it askand micro levedesign of the sp

individuals enrstand and mea

2019

d the macro ues to

work ins in bber, ty of ers of

two xemic ssical users gton,

bber, earch s are

macro t also unity

ndoor hear,

ulture ng of aphic

ks the This

s the els. It paces

ngage asure

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

34

what was impacting interactions or engagements. We believed it was important to not just answer the research question, but try and provide a tool, or index and measurable awareness levels to use as gauges of engagement and environmental fit. This document reports on a survey of three middle schools, grades 6–8, and is the first time we have attempted to survey students in this age group, as an ‘alpha’ test, and expected to learn some things to do differently the next time around. Our earlier surveys were of high schools, grades 9–12. Surveying middle school students raised several questions:

1) Would middle school students understand a survey of the type we need to do? 2) How much language would need to change to ensure these cohorts would comprehend our questions from

high school? 3) Would middle school students be willing to respond to a survey? 4) How much would the survey need to be simplified and/or shortened for them? 5) Would we get results similar to the high school surveys, and how would results differ? 6) How might middle school teachers answer differently than high school teachers?

We are pleased that overall, students responded well to the survey, and the results were similar to those of the high schools’ surveys. To address our concern stated in #2, we asked a former assistant superintendent and some of her educators to review the 9–12 text and help ensure the vocabulary and meaning would fit with this age cohort. Some slight changes were made to the original surveys (three pilots for high school) as a result of their reviews. While we encountered a couple issues along the way, there was nothing that would call the validity of the survey into question. As one would expect, there were some differences between the middle school and the high school survey responses. 2. Method A short definition for each step in this Human-Centered Research Design protocol (refer back to Figure 1) is shared:

• Discover [D]: Develop a research question/hypothesis and understand what will be the best research design, methods and techniques to find answers, and use them to gather data. It’s best to use three techniques to ensure bias is reduced. Once gathered the researcher(s) puts this information into appropriate format(s) for analysis. Whether using quantitative methods or qualitative methods, all data will be worked to produce some numerical findings. Once this latter stage is done, these become research ‘instruments’ or tools. (NOTE: a human subject’s protocol has been reviewed by a third party prior to beginning work with a client; all consent forms approved and received).

• Analysis [A]: Take the data from the research techniques and use appropriate methods to break down the information. By using multiple discovery techniques to avoid bias ensures the comparisons generates consistent and reliable findings. Use statistical methods when appropriate. Pilot test and test again to vet the data for reliability and validity.

• Synthesis [Sy]: Recognize the analysis phase of this work only generates facts. What these facts ‘are saying,’ how each is connected to the next is revealed by generating meaning and understanding relative to the original question. This segment takes time and expertise to clarify and built a ‘truthful’ and unbiased consensus from the data.

• Share [S]: Be prepared to share information to multiple audiences and for multiple purposes—clients, designers, conferences, and research manuscripts.

• Plan [P]: Know all data reveal a truth—not always the ones we’re looking for or expecting. Be prepared to plan for next steps (ex, go back and address an issue found in a design and/or adjust the design solution for the next time it is used).

The sample information and response rates are shared next. 2.1 Sample and Response Rates and Analysis of Missing Data There were three schools used as convenience samples with purposeful user groups (students and teachers), each school designed by a particular architectural firm, with 1,381 total responses to the student survey. Student response to the survey was very good, with over 60% of the students in each school responding to the survey. Response rates from the teachers were much lower at schools A and B; the reason is unknown (see Table 1). The small number of teachers responding made it difficult to draw statistically valid inferences about how the

jel.ccsenet.o

teachers vhere are fr Table 1. R

We had a Figure 3).

Similarly, more than

Of the “Otlittle probland anothefatigue” in5).

org

viewed the diffrom the 6-8 stu

Response rates School

A B C

good mix of s

we had a gooone grade; see

ther” group, thlem with non-rer 12% skippen the student su

ferent schoolsudy.

by school No. Students

848 479 680

students by gr

Fi

od mix of teace self-identifie

Fig

here was one aresponse to qued only one quurvey, as the n

Journal of Ed

. To ensure an

Respondents 545 403 433

rade level, whi

igure 3. Grade

chers by gradeed other grade

gure 4. Grade l

answer each ouestions on theuestion (but nonumber of skip

ducation and Le

35

nonymity, sch

% No64.3 84.1 63.7

ich proved to

level indicato

level; focus olevels taught (

level indicator

f “Office”, “Ge survey. Amonot the same qupped questions

earning

ool names are

o. Teachers R120 45 45

be an importa

rs/students

on grades 6-8.(see Figure 4).

rs / teachers

Grade 5”, and “ng the studentuestion). Hows rose toward t

e coded, and a

Respondents 31 20 29

ant factor in th

. Of course, m

“Counseling”. ts, 68% answer

wever, we did the end of the

Vol. 8, No. 5;

all data represe

% 25.8 44.4 64.4

he student data

many teachers t

Overall, therered every quessee a little “susurvey (see F

2019

ented

a (see

teach

e was stion, urvey igure

jel.ccsenet.o

Eighty teaLooking anumeric-raquestions w95% or beshown in tsupported

3. Results3.1 ReliabThe objectquestions Alpha is away. We hsurvey, onrating valuwith only and the stuThe idea oAnother mare strongevidence othese survdifference surveys, an

org

achers in all tooat the questionating item, andwas not a probetter on those the graph belothe teaching m

ility & Validitytive behind meon the survey

a number betwehave very goodnly question 4 ues were both gquestion 1 (im

udent surveys aof validity in a

method to detergly correlated of validity in beys are similais that the re

n intuitively p

Figure 5.

ok the survey, n items with d 4 teachers skblem on this suquestions. Fivw. The greates

method the teac

Figure 6.

y easuring reliaby. Consistent ween 0 and 1, wd reliability nu(classroom angreater than .7

mportance of vare reliable. a survey is to rmine whetheras expected)

both the studenar to, but not idesponse by grapleasing result,

Journal of Ed

Survey respon

although one numeric answkipped only onurvey; for eachve teachers quist non-responscher used (see

Survey respon

bility of a survewith the earlie

with a value of umbers overallnd building rat70. On the studvarious items)

ask whether r the survey is

and divergennt and the teachdentical with, ade level is of, given the yo

ducation and Le

36

nse rate—surve

person answewers (1–4 and

ne item (see fh question grouit after questiose was for the Figure 6).

nse rate—surve

ey is to assess er surveys, wef 1 indicating thl for both the ttings) had Alp

dent survey, allbeing less tha

the study meas providing thent (weak or nher surveys. Athe surveys d

f greater impoounger ages of

earning

ey fatigue/stud

red only the d6–9), 70 of

figure 6). Non-up with numer

on 9, so there second part of

ey fatigue/teac

whether peope used Cronbahat questions ateacher and thepha below .80,l the values of

an .80. Thus, w

asures what the results intendnegative corre

A strong indicadone of high scortance in thisf the students.

dents

demographic quthe 80 teache-Response of ric answers thewas a bit of “f question 10,

chers

le give similarach’s Alpha toare answered ie student surve, and the classf Alpha were awe conclude th

he researcher dded is throughelation) validitator is the fact chool studentss survey than Teachers also

Vol. 8, No. 5;

uestions at theers answered eindividual nume response rate

“survey fatiguewhether the de

r answers to simo assess reliabin exactly the eys. On the teasroom and builso greater tha

hat both the tea

desired to meah convergent (ity. We have that the findin

s and teachers.in the high sc

o seem to sens

2019

end. every meric e was e”, as esign

milar bility. same acher lding n .70, acher

asure. items good gs of One chool e the

jel.ccsenet.o

building asConvergenvariables; theoreticaland low coA similar pand questiinterest, al3.2 StatistiClassroom1 as beingteachers anseems to bstudents wtemperaturon our 5-pwere mino“mentoringsociability(testing, cr

org

s having a greant and divergequestions 2–9l question 1 (Iorrelations whepattern occurs ons 2, 3, 4, 6, long with a comics and Data A

m ratings overag more importnd students. Tbe universal buwere slightly mre “Excellent”point scale. Thor. The most dg” shows less

y, i.e., inter-perritical thinking

F

Fig

ater impact onent validity in, about how thImportance ofere one wouldin the teacher 8, and 9 have

mparison of stAnalysis all were fairly tant than did t

The environmeut the studentmore critical

” or “Very goohe average studdivergent varias divergence. rsonal relation

g, creativity, et

Figure 7. Impor

gure 8. Environ

Journal of Ed

n them than at tn the student hings actually w…) has lower

d expect them tr survey. Quest

good correlatitudents and tea

similar (see Fithe students. Q

ental quality qus rated the nothan the teac

od”. The overadent rating for ables are “beloThese three m

nships and engtc.).

rtance of items

nmental quality

ducation and Le

37

the high schoosurvey may b

work and are wcorrelations w

to be. tion 1 has relations with eachachers are offer

igure 7). Not sQuestion 1 hauestions generise level as behers. For both

all average ratir Noise Level wonging to the more than the gagements, rat

s for engageme

y ratings from

earning

ol level. be seen in thwell correlatedwith all the oth

tively low corrh other. Some ired next.

surprisingly, tead the greatestrated complaineing just as bah students aning by studentswas 2.74 (see school’s commother six var

ther than indiv

ent/students vs

both student a

STU

DEN

TS

STU

DEN

TS

he correlationsd with each othher variables.

relations with individual que

eachers saw tht divergence onts about the tad (see Figure nd teachers, ons was 2.78, anFigure 9). Difmunity” and “riables are relavidual experien

s. teachers

and teachers

TEAC

HERS

TEAC

HERS

Vol. 8, No. 5;

s of the compher, while the Thus, we see

the other variaestions of parti

he items in queof opinion betwtemperature, w8). In genera

nly 26% ratednd 2.74 for teacfferences by sc“collaboration,ated to inter-ances and outco

2019

posite more high

ables, cular

estion ween

which l, the d the chers chool ” yet

active omes

jel.ccsenet.o

On questio

3.3 The RoSome demtheir gradethey held. French, Lschool levlevel prodschool. In the student3.4 PrincipBefore coquestion gquestion gFor each nwhole grouset to a miThe compo

org

on 6, the value

ole of Demogramographic infoe level, and wWe knew whiembke, & Kin

vel, the schoolsduced some int

the regressiont grade level mpal Componen

onstructing thegroup, and a pgroup would benumeric questup. If more thaissing value. Tosite variable f

Figure 9. Per

es of the school

Figure

aphic Factors rmation was c

we asked the teich school theynney, 2017; Ss in this surveyteresting diffens described bmattered. nt Analysis & Ce composite vprincipal compe a reasonable tion group, thean one questiohe composite vfor question 7

Journal of Ed

rceptions of no

l, teachers wer

e 10. Perceptio

collected on theachers what gy were at withoScott-Webber, y did not becorences, perhap

below, demogr

Composite Varvariables, descponents analysproxy for the g

e mean of the on in the groupvariables were(At the end of

ducation and Le

38

oise and tempe

re sometimes m

ons of what the

he surveys. Forgrades they tauout asking. UnKonyndyk, F

ome a factor inps a reflectionraphic factors w

riables cribed below,sis was perforgroup. group was us

p was skipped, e used in the ref a school day,

earning

erature comfor

more positive t

e school values

r the students ught, their gennlike the “OmeFrench, & Frenn most of the qn of the changewere checked

, Cronbach’s rmed in order

sed as the comthen the comp

egression analy how often do

STU

DEN

TS

rt levels

than the studen

s

we asked abounder, and whatega’s” (Scott-Wnch, 2018), su

questions. For es a child expfor relevance,

Alpha was cto verify that

mposite variabposite variableyses and in theyou feel that y

TEAC

HERS

Vol. 8, No. 5;

nts (see Figure

ut their gendert academic degWebber, Konyurvey for the students, the g

periences in m, but typically

computed for t the mean of

ble to represene for that groupe cluster analysyou…) was us

2019

e 10).

r and grees

yndyk, high

grade middle

only

each each

nt the p was sis. ed as

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

39

the engagement index for both teachers and students. The overall average values for both the teacher engagement index (TEI) and the student engagement index (SEI) were similar to the corresponding values on the “Omega” survey, and like that survey, the student values showed greater variability (see Figure 11). The results of the analysis are next.

Figure 11. Distributions of the engagement indexes

3.5 Analysis Results 3.5.1 Regression Analysis/Students In general, the school played little or no role in the regression results in this survey, while gender had some impact, and grade level had even more impact. Consistent with the surveys of high school students, girls tended to have slightly higher engagement than boys, and those who opted for the “prefer not to say” sector having lower engagement. Engagement also tends to drop as the grade level rises. Results of the regressions in this survey were very similar to those of the “Omega” at the high school level. While there were a couple of statistically significant interactions, they were of little practical importance. The impact of the various questions on student engagement is basically the same across school, gender, and grade level. These are shared next:

• SEI and Question 1: (The importance of various items for engagement). Of all the numeric questions (1–4, 6, 8, and 9), question 1 has the weakest relationship with student engagement. Regressing the SEI on just question 1 yields a small r2 = .13. Students who perceive that the items in question 1 are important show only a slight tendency to have a higher level of engagement.

• SEI and Question 2: (How well do the classrooms provide you with the ability to…?). Regressing the SEI on question 2 by itself gives a very strong r2 = .29 with p < .0001. The school is not a factor, but adding in grade level and gender to the regression raises r2 to .34. Also, there is no interaction of Q2 with either grade level or gender, and we see that students who believe that the classrooms provide the ability to see and hear well and give access to appropriate items are also likely to be more engaged. The values of r2 here are quite similar to those of the “Omega” survey (see Figure 12).

Student Survey

Mean: 3.73

Std Dev: 0.82

Teacher Survey

Mean: 3.98

Std Dev: 0.46

STUDENT TEACHER

jel.ccsenet.o

• SE

qleinstQ

• SE

asraanleef

org

SEI and Questquestion, the scevel, though nnteraction betwtatistically sign

Q3 to predict th

Figu

SEI and Questisked to rate thate the buildinnd engagemenevel, though stffects (see Fig

Figu

tion 3, (How mchool figures innone with schoween question nificant, only he SEI gives r2

ure 13. Regress

ion 4: (Pleaseheir classroomsng overall, a cnt was fairly sttatistically sigure 14).

1

1.5

2

2.5

3

3.5

4

4.5

5

Q

Journal of Ed

ure 12. Regres

much impact dnto the regressool or gender.

3 and grade lincreases r2 fr

2 = .26, with p

sion plot of the

e rate the follos on several aschange from thtrong, with r2 =nificant, raised

Q2 How well c

ducation and Le

40

sion plot of th

does the desigsion, and there Thus, using qlevel yields r2

om .30 to .31,< .0001, show

e SEI on Q3 (im

owing aspects spects (noise, the grades 9–1= .29, very simd the value of

classrooms prto...

earning

he SEI vs. Q2

gn of the class is an interacti

question 3, gen2 = .31, with p, resulting in li

ws a good assoc

mpact of class

of the classrotemperature, li2 survey. The

milar to the “Of r2 by only .0

rovide the abi

sroom have oion between qunder, grade levp < .0001. Theittle practical iciation by itsel

sroom design)

oom environmeighting, etc.), te association b

Omega” survey3, indicating t

ility

Vol. 8, No. 5;

on you…). Foruestion 3 and gvel, school, ane interaction, wimpact. Using lf (see Figure 1

ent). Students though they dibetween the ray. Gender and gthat they are m

2019

r this grade d the while only

13).

were d not

atings grade minor

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

41

Figure 14. Regression plot: SEI regressed on Q4 (classroom ratings)

• SEI and Question 6: (What do you think your school values…). Consistent with the “Omega” survey of

grades 9–12, we see a strong relationship between student engagement and whether the student believes that the school values items such as creativity, critical thinking, and mentally challenging work. Regressing the SEI on question 6 alone gives r2 = .39. (The corresponding regression on the “Omega” survey had a similar r2 of .45). This question and question 8 (impact of the design of the school on access to peers and teachers and on the ability to move in the classroom) had the strongest association with student engagement, of all the questions.

Once again, school is not statistically significant, even at the p = .05 level. Adding grade level and gender into the regression gives only a small improvement in r2, to .42. We may conclude that the effect of the values of the school is the same across gender, grade level, and school, and the effect of the perceived values on engagement is quite strong (see Figure 15).

Figure 15. Regression plot: SEI vs. values of the school

• SEI and Question 8: (How much to you believe the design of the school overall impacts your ability)

to….). Once again, school is not significant, nor is any interaction terms. Grade and gender add only a little to the information provided by question 8, improving r2 from .42 to .45. We have another very

1

1.5

2

2.5

3

3.5

4

4.5

5

Q4 Ratings of Classrooms

1

1.5

2

2.5

3

3.5

4

4.5

5

Q6 Values of the school

jel.ccsenet.o

stte“Osi

• SE

reinst

In summarbuilding an3.5.2 Regr3.5.2.1 ThLooking a(TEI). Thea lower mone-way Alevel of edthe p = .00(see Figure

org

trong associatieacher and proOmega” (9–12imilar r2 = .39

SEI and Questelationship betn the grade levtatistically signry, these regrend the student ression Analys

he TEI and Demat gender by itse means of the

mean of 3.58, bANOVA for thducation appea012 level, givine 17).

ion here; studeovides the ab2) survey, the(see Figure 16

tion 9: (Leveltween Q9 and vel improves nificant, but ofessions show aengagement in

sis/Teachers mographic Varself, there weree TEI were an but with only the TEI on levar to have no ng r2 = .166, w

Figure 17.

1

1.5

2

2.5

3

3.5

4

4.5

5

Q

Journal of Ed

ents who believility to move e regression o6).

Figure 16. Re

l of impact ofthe SEI, with r2 to .33. Genf little practicaa strong statisndex.

riables e no detectableall-but-identicthree people in

vel of educatioimpact on the

with school C h

Boxplot of tea

Q8 Impact of

ducation and Le

42

ve that the desin the classro

of the SEI on

egression plot

f the design or2 = .30 with

nder, school, aal importance, tically signific

e differences ical 3.98 for won that group, t

on yielded no e TEI. Differenhaving the high

acher engagem

design of the

earning

sign of the schooom tend to hthe correspon

: SEI vs. Q8

of the buildingthe SEI regres

and the interacas each of themcant relationsh

in the values oomen and 4.01there is no stastatistically signces by schoolhest mean (4.1

ment index by s

e school over

ool facilitates have higher ending question

g on your…).ssed on questioction of grade m improves r2 hip between sa

of the Teacher for men. “Pre

atistical signifignificant diffel were statistic

17) and school

school

rall

Vol. 8, No. 5;

access to peerengagement. Inn produced a

. We see a ston 9 alone. Ad level with Qby only .01.

atisfaction with

Engagement Iefer not to say”icance. Similarerence. Gendercally significan A the lowest

2019

s and n the very

trong dding 9 are

h the

Index ” had rly, a r and nt, at (3.75)

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

43

3.5.2.2 The TEI and the Numeric Questions (1–4 and 6–9) Perhaps because of the smaller sample size in this survey, the relationships between the numeric questions and the TEI lack statistical significance. As we saw with the “Omega” 9–12 survey, the school is a better predictor of the TEI than are any of the numeric questions. A summary of this information is next:

• Question 1, the second part of question 1 (Design supports), and question 2 do not come close to statistical significance for the TEI, whether with or without the demographic variables

• Question 3 hovers near statistical significance, but only at the p = .05 level. The school is of greater importance to the TEI than is question 3. Regressing the TEI on just question 3 yields a measly r2 = .055, with p = .0425, while including the school in the model gives r2 = .22.

• Question 4 Classroom Ratings are statistically significant, with r2 = .10 and p = .0052. Again, school is more important; adding it to the model raises r2 to .25. This actually a bit stronger relationship than we saw in the “Omega” survey.

These results are similar to those of the “Omega” from grades 9–12 survey, which also found only weak relationships between teacher engagement and the other numeric questions. Cluster analysis is described next. 3.6 Custer Analysis 3.6.1 Clusters for the Students For the students, four clusters worked well. As with the previous surveys, Ward’s Method was used for the clustering. Average engagement in the clusters varied widely, from a high of 4.70 in cluster 2 to a low of 2.91 in cluster 4. Alas, cluster 2, with the highest average engagement, is the smallest cluster. Except for question 1, part 2, a higher number means “good” and a lower number means “bad” for the questions. Question 1, part 2 asked whether the design helped the items in the first part of Q1 succeed, and the scale of question 1, part 2 was reversed, with 1 = Yes and 2 = No. That explains the strange look in the graph at that question. Thus, the four clusters are quite neatly stratified, with cluster 2 having the highest means for each question, followed by cluster 1, and then cluster 3, with cluster 4 having the lowest means for each question (see Table 2). Table 2. Student cluster means

Cluster Count Q7 Student Engagement

Index

Q1 Importance of various items

Q1 part 2 Is the design helpful? (1

= Yes, 2 = No)

Q2 How well classrooms provide

the ability to...

Q3 Impact of classroom

design

Q4 Ratings of

Classrooms1 356 4.19 3.57 1.06 4.15 3.50 3.62 2 133 4.70 3.97 1.03 4.51 4.58 4.08 3 499 3.46 3.20 1.14 3.43 2.73 2.91 4 182 2.91 2.66 1.55 2.85 2.16 2.38

Cluster Q6 Values of the school Q8 Impact of design of the

school overall Q9 Impact of design of building's physical spaces

1 3.97 4.14 3.70 2 4.57 4.76 4.81 3 3.22 3.17 2.82 4 2.48 2.40 2.10

A corresponding graph of the cluster means is in Figure 18 (see Figure 18).

jel.ccsenet.o

A look at tFigure 19)

Consistentand 2, with

org

Figure 18

the clusters by).

t with the regrh a bigger drop

8. Correspond

y school does n

ression results,p-off in “prefe

Journal of Ed

ing cluster gra

not show stron

Figure 19. Cl

, one can see aer not to say” w

ducation and Le

44

aph of interpret

ng differences,

uster means by

a little higher when viewing c

earning

tation of desig

, a change from

y school

percentage of cluster by gend

Studen

gn support/stud

m the “Omega

f females than der (see Figure

nt

Vol. 8, No. 5;

dent

a” 9–12 survey

males in cluste 20).

2019

y (see

ters 1

jel.ccsenet.o

The graphdirection!’

3.6.2 ClustThe results3 (blue) anphysical eengaged tvarious qu

org

h of clusters by’ Next, is the c

ters for the Tes of the clusternd 4 (orange) benvironment, teachers helps

uestions (see Fi

y grade level clustering for th

achers r analysis of thboth show higthe 9 teachers

explains whyigure 22).

Journal of Ed

Figure 20.

(Figure 21) ishe teachers.

Figure 21. C

he teachers aregh engagements of cluster 3 y the regressi

ducation and Le

45

. Cluster by ge

s more dramat

Cluster by grad

e reminiscent ot, while the 26are quite unh

ons showed s

earning

ender

tic, and the ind

de level

of the surveys teachers of cl

happy with it.so little conne

dicators are go

of high schooluster 4 are ver This diverge

ection between

Vol. 8, No. 5;

oing ‘in the w

ol teachers. Clury happy with

ence among hn the TEI and

2019

wrong

usters their ighly d the

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

46

Figure 22. Corresponding cluster graph of interpretation of design support/teachers

Table 3. Cluster analysis

Cluster Count Q7 Teacher Engagement Index

Q1 Importance of various items

Q1 Design supports for activities

Q2 How well classrooms provide the ability to…

Q3 Impact of classroom design

1 23 3.908 4.061 1.087 3.555 3.065 2 16 3.758 2.900 1.238 3.392 2.292 3 9 4.097 4.333 1.956 2.374 2.870 4 26 4.144 4.046 1.015 4.024 3.494

Cluster Q4 Ratings of

Classrooms Q4 Rate the building overall

Q6 Values of the school

Q8 Impact of design of the school overall

Q9 Impact of design of building's physical spaces

1 3.060 3.207 3.588 3.661 3.435 2 3.102 3.188 3.083 3.125 2.696 3 3.083 3.208 3.370 2.667 3.079 4 3.928 4.024 4.175 4.208 4.203

The cluster means are in the Table 3 (refer back to Table 3). As with the student survey, the second part of question 1 (“Design supports?”) had 1 = Yes and 2 = No. Therefore, low values are desirable rather than high values for that question. Note that cluster 4, with the highest TEI, also had the “best” mean for each question other than the first part of question 1. Cluster 3, which also had a high mean TEI, had the “worst” ratings for the “Design Supports” part of Q1, and questions 2 and 8, along with nearly bottom ratings for questions 4 (both parts) and 6. Clusters 1 (red) and 2 (green) had similar average TEI values, but cluster 1 assigned very low importance to the items in question 1, while saying that they were very well supported, and cluster 1 saw much more impact from the building than cluster 2 (questions 3, 8, and 9) Comparing the teacher cluster by school, results show that School B has the highest percentage of teachers in cluster 4, the “happy” group, with the highest means for TEI, and no teachers in cluster 3, the group that seems quite dissatisfied with their classrooms, despite having a high TEI (see Figure 23). What we found regarding the impact the physical surroundings has on the teachers and students follows.

Teacher Cluster 1

Cluster 2 Cluster 3

Cluster 4

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

47

Figure 23. Clusters by School

4. Discussion 4.1 Impact of the Physical Surroundings 4.1.1 Teachers The table below shows the results of t-tests for H0: Mean = 3 vs. Mean > 3 for the questions about the impact of the building (see Table 4). Meanings of the numbers:

• Q3: 3 = Moderate impact - 4 and 5 are Makes a noticeable (or big) difference. • Q8: 3 = Acceptable - 4 and 5 are Easy or Very easy. • Q9: 3 = Moderate impact - 4 and 5 are Makes a noticeable (or big) difference.

Different question items elicited distinctly different average values of the perceived impact of the built environment. Items significant at the p = .01 level are underlined. The overall view seems to be that the impact of the design of the school on teachers’ ability to do things is better than merely “Acceptable” (Q8), but that the impact of the building on teachers (Q9) is greater than the impact on students (Q3). Average values in question 3 are slightly lower than in the “Omega,” while answers to Question 9 tended to be a little higher than for the corresponding questions in the “Omega.” Table 4. Impact of built environments/teachers (selected items)

Question Item N Mean T-Test statistic P

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

48

4.1.2 Students Students were asked most of the same questions as the teachers about the impact of the built environment. Question 3 asked about the impact of classroom design on the student’s attitudes. Question 8 asked about the impact of design on the student’s ability to do certain activities, and question 9 asked about the impact of the physical spaces on various things. The table below shows the results of t-tests for means of 3 or higher (see Table 5). For questions 3 and 9, this means at least a moderate impact vs. a little or no impact. For question 8, this means the impact is acceptable or better. Overall, the averages here are a little higher than those on the “Omega” survey. The descriptors by grade level are next. Table 5. Impact of built environment/students (selected items)

Question Item N Mean T-Test statistic P

jel.ccsenet.o

4.1.4 “Mov4.1.4.1 StuComparingreported ahand-in-hamove, is aquestion 1actually be

org

Figu

vement” Quesudents/Movemg “movementabove, but theand with highea bit theoretica1 are less stroeing able to mo

ure 24. Grade l

Figure 25.

stions ment t” questions we point is woer student eng

al, and not surpongly related tove, as shown

Journal of Ed

level of studen

. Averages Q6

with student orth repeating gagement. For prising that quto engagemenin the table be

ducation and Le

49

nt vs. mean val

: what the scho

engagement sfor some indstudents, quesestion items ab

nt than are theelow (see Tabl

Y

earning

lues of compos

ool values/stud

somewhat dupdividual questistion 1, about bout the percee questions itele 6).

YR 6

YR 7

YR 8

site variables

dents

plicates the reions. The abithe importanc

eived importanems in questio

Vol. 8, No. 5;

egression anaility to move ce of being ab

nce of movemeons 8 and 9 a

2019

alyses goes

ble to ent in about

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

50

Table 6. Questions related to movement and engagement Question Item Correlation with the SEI Q1b Move about classroom to be actively engaged - Importance of… 0.248 Q1d Have the choice to use different parts of the room to work with others - Importance of… 0.208 Q8c Ability to move to engage in classroom activities 0.573 Q8d Choose either to sit/lounge/stand in order to be active in the classroom 0.496 Q9c Ability to move around to become deeply engaged in my learning 0.507

The overall message is that for these middle school students, the ability to move correlates strongly with student engagement. A look at the “ability to move” question items in question groups 8 and 9, and engagement shows that engagement is higher for students who believe that movement is easier (see Figure 26). (A note about boxplots: the line inside the box indicates the median value, and the box itself contains the middle 50% of the values, giving an idea of the “spread” of the values.)

Figure 26. Ability to move corresponding to level of engagement / students

4.1.4.2 Teachers Teachers were also asked some questions directly relating to movement, specifically:

• Q1a Transition in and out of small groups—Importance of… • Q1b Move about classroom to be actively engaged—Importance of... • Q1d Have the choice to use different parts of the room to work with others—Importance of... • Q2i Move around to keep students engaged (How well do classrooms provide this ability) • Q2j Have your students move around to keep themselves engaged • Q8c Impact of the design of the school on: Ability to have your students move to engage in classroom

activities • Q9c Impact of the physical space on: Ability for you to move around to get your students deeply

engaged in their learning. While the student answers to the analogous questions in their survey were positively correlated with the student engagement index, the teachers’ answers to their questions showed little relationship to teacher engagement. The questions listed above are all positively correlated with each other, and they fall into two basic groups: the three parts of question 1 are well-correlated, indicating that teachers who viewed one type of movement as important tended to view the other types of movement as important. The second group consists of items 2i, 2j, and 8c,

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Q8c Ability to move to engage inclassroom activities

Q8d Choose either to sit/lounge/stand in order to be active in the

classroom

Q9c Ability to move around tobecome deeply engaged in my

learning

jel.ccsenet.o

which havitems in qgrouped wthey believability to m

Table 7. M

4.1.5 TeacAccordingdone outsistudents dschool A. teachers (s

Students restudents ge

org

ve to do with question groupwith the items ive that movemmove (see Tab

Mean answers t Q1a TransitionQ1b Move aboQ1c ExperiencQ1d Have the Q2i Move arouQ2j Have yourQ8c Ability toQ9c Ability fo

ching Outside tg to the teacheride the classroo not completStill, student

see Figure 27).

eport little diffet outside of th

the ability to p 1. Question n question 1. T

ment is importle 7).

to “movement”

n in and out of smout classroom to bce hands-on activichoice to use diffund to keep studer students move a

o have your studenor you to move aro

the Classroomrs, very little t

oom at school tely agree withs in school B.

ference in nonhe classroom m

Journal of Ed

move, and theitem 9c is on The means of ttant (question

” / teachers

mall groups (of 2 tobe actively engageities during class tferent parts of the nts engaged round to keep them

nts move to engagound to get your s

m teaching is donB; no teacher

h the teachers, B seem to repo

Figure 27. Te

-classroom teamore (see Figu

ducation and Le

51

ey are stronglyits own, thou

the teachers’ a1), but are pe

o 4) - Importance ed - Importance oftime - Importanceroom to work wit

mselves engagedge in classroom actstudents deeply en

ne outside the r at school B rin that they re

ort the most t

eaching outside

aching by gradure 28).

earning

y correlated wugh a factor ananswers to the rhaps a little l

of... f... e of... th others - Importa

tivities ngaged in their lear

classroom at sreported alwayeported much teaching outsid

e the classroom

de level. Intere

with each othenalysis sugges“movement” qless enthusiast

Me3.73.93.8

ance of... 3.63.43.43.6

rning 3.7

school A, whilys staying in tmore non-cla

de the classro

m

stingly, engag

Vol. 8, No. 5;

er, but not withsts that it coulquestions showtic about the a

ean 78 96 88 67 49 41 62 76

le more teachithe classroom.ssroom teachin

oom, similar to

gement rises a b

2019

h the ld be

w that actual

ing is . The ng in o the

bit as

jel.ccsenet.o

A one-wayalways taufor the “mmethods aperspectiv4.1.6 Teac4.1.6.1 StuThe studenprojects sestudents re

org

y ANOVA shught in the clamore than halfare often a prves. ching Methodsudent Responsnts reported oneem to decreaseporting each m

Figure 2

hows that the dssroom, the mf” group. Thesredictor of act

and Strategieses nly minor diffese, while lecturmethod as one

Figure 3

Journal of Ed

28. Teaching o

differences in mean engagemese results are tive engageme

s Used

erences by schre and discuss of the top two

30. Teaching m

ducation and Le

52

outside the clas

means are reent is 3.58, cosimilar to wh

ent by the stu

hool, but as theion increase (s

o methods used

methods by gra

earning

ssroom / studen

al (F = 8.0289ompared to 3.9hat we saw inudents. These

e grade level rsee Figure 30)d.

ade level/stude

nts

9, with p < .091 for the 6–25n the “Omega”

data are shar

rises, hands-on. The graph sh

nts

Vol. 8, No. 5;

0001). For stud5% group and ” survey. Teacred from both

n learning and hows the perce

2019

dents 3.94

ching user

team ent of

jel.ccsenet.o

Another im

4.1.6.2 TeaIn questionso only theput it neartake-homedifference

There werQuestion 1of the stud

org

mportant way t

acher Responsn 10, which ase total groupinr the top; an obe messages; te

to student out

re too few resp12 (Teaching sdents (see Figu

to illustrate the

F

ses sked about teacng is shown hebvious and imeaching stratetcomes. Perhap

F

ponses to detestrategies used

ure 33).

Figu

Journal of Ed

ese perception

Figure 31. Tea

ching methodsere. The teache

mportant percepgies and pracps this data sho

Figure 32. Teac

ect differencesd), differences

ure 33. Teachin

ducation and Le

53

s from student

aching methods

s used, differeners put lecture/ption to considctices to ensuows a best prac

ching methods

in the “Desigby school wer

ng strategies u

earning

ts is in the figu

s/students

nces between /presentation nder (see Figureure a variety octice scenario

/ teachers

gn Supports?” re not signific

used/teachers

ure below (see

schools were nnear the bottome 32). One of of activities tfor considerati

answers by teant. This ques

Vol. 8, No. 5;

Figure 31).

not significantm, while the stuthe most impotake place maion.

eaching methostion was not a

2019

t, and udent ortant ake a

od. In asked

jel.ccsenet.o

In the “Deof the 37 f

5. LimitatAction resresponses,determinindesign of tFirst, whilstudent enexample, twith the bsample of 6. ConclusOverall, thsurveying seemed ot high schomiddle schinteraction

•

•

•

•

•

•

org

esign Supportsfor “Project-ba

tions search at the p and therefore

ng the types othe learning plle the survey

ngagement, it mthat an overalbuilding and cthree schools tsion his survey “wo

middle schoounderstand th

ol students, shool students, n. Some specif

Students aengagemenTeaching tstudents’ alearning s(BehavioraMoving torecognizedConnectingoutside walevels of enWhen respengagemenlearning ouAs with tengagemenwas much

s?” portion of qased, and 2 of t

Figure 34

post-occupance controlling foof teaching strlace afforded ths showed higmust be noted l positive attitclassrooms. Seto all schools i

orked”, althouols, the surveyhe questions asatisfaction wiand the corre

fic findings areand teachers bnt levels at a hitechniques chattitudes seemtrategies, the

al Responses) o learn was imd its importancg to nature anas not done verngagement (Inpondents recognt, the perceptutcomes, and mthe high schont and satisfactweaker (Envi

Journal of Ed

question 12, othe 29 answers

4. Teaching str

cy level recognor variables is rategies used vhose activities

gh correlationsthat correlatio

tude could be econd, one muin the country.

gh some tweay went well. Ts they answereith the physicespondence is e shared in termboth perceivedigh rate of sign

hanged with gm to become

more their

mportant as ie (Environmen

nd being outsiry often. Whenternal Respons

gnized the valuion was there

motivation to aool surveys, tion with the pironmental Dim

ducation and Le

54

nly 7 of the 13s of “Personal

rategy & does

nizes that therchallenging. T

versus the lev. s between saton was not the

behind both bust be very ca.

aks would be hThe students sed like those oal facilities gthe same acro

ms of how theyd the design onificance (Envgrade level, amore negativstudents elici

t increased enntal Dimensionide to learn wn it was, studeses) ues of the schwere higher l

attend classes (we found strphysical enviromensions). Fo

earning

33 respondentLearning” (see

design suppor

re are multiplThis survey wvel of student

isfaction withe same as causbeing more enautious in gen

helpful. Given showed only af the high schooes hand-in-hoss gender, gry connect to thof the built spvironmental Diand along wite. Thus, the ited higher le

ngagement levns)

was viewed as ents responded

hool, or its situlevels of stude(Value Dimensrong correlationment, howevr teachers, the

s gave an answe Figure 34).

rt it?

le factors that worked to inclu

activity perm

h the physicalsation. Someonngaged and beneralizing from

that this was a little ‘surveyool students. A

hand with engade level, and

he UEIF.v2 arepace impactedimensions) th these moremore teacherevels of acad

vels, and both

important, bud positively tha

uational culturent academic esions) ons between ver, for teachee cultural/situa

Vol. 8, No. 5;

wer of “Not w

contribute to ude some contr

mitted, and how

l surroundingsne might argueeing more satim our non-ran

the first attemy fatigue’ and As was the casgagement for d school, with e: d student acad

e didactic chas embraced ademic engage

h response gr

ut actually moat it increased

re relative to aengagement, h

student acaders, that conneational climate

2019

ell, 5

user rol in w the

s and e, for isfied ndom

mpt at they

se for these little

demic

anges active ment

roups

oving their

active igher

demic ction

e was

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

55

more important (Value Dimensions). Both student and educator perspectives are important to provide a holistic understanding of how the design of a space for learning impacts student outcomes. A high level of statistical significance has been reached across all schools, geographies participating, grade levels, demographics and user groups that YES the design of space makes a difference for student academic engagement levels at p < .0001. The knowledge gained across the level of cohorts we measured, and will continue to measure, impacts the way one should consider developing design solutions from the macro level to the micro level. The UEIF.v2 continues to ensure questions connect to this grounding framework. The 21st century learning goals, and teaching strategies to support them are well articulated, and through post-occupancy analysis we bridged the connection between design and performance. References Abonyi, J., & Feil, B. (2007). Cluster analysis for data mining and system identification. Boston, MA:

Birkhauser Basel. Anonymous. (2010). Working with NSSEdData: A facilitator’s guide. National Survey of Student Engagement.

Retrieved from http://www.nsse.iub.edu/links/facilitators_guide Anonymous. (2008, April). Implementation of post-occupancy evaluation: A potential tool for building asset

management and creating more productive, cost-effective and sustainable buildings at MSU (pp. 1–7). White Paper # C2P2AI, SPDC. Michigan State University.

French, J., Scott-Webber, L., Ferking, T., & Fulton, D. (2015). Early childhood space design: Nurturing every student’s potential (p. 20). EDspaces Conference. LA: education Market Association.

Hanington, B. M. (2010). Massachusetts Institute of Technology Design Issues, 26(3), 22. Retrieved August 13, 2019 from https://s3.amazonaws.com/academia.edu.documents/31490953/DESI2603_Hanington_pp018-pp026_v4.pdf?response-content-disposition=inline%3B%20filename%3DRelevant_and_Rigorous_Human-Centered_Res.pdf&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIWOWYYGZ2Y53UL3A%2F20190813%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20190813T185400Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&X-Amz-Signature=73057dc18edb4d8951f04d7c8722947fae5f39e25bbe6020811d68c9eff48b56

Johnson, BR, Onwuegbuzie, AJ, & Turner, LA. (2007). Toward a definition of mixed methods research. Journal of Mixed Methods Research. Vol. 1, pp. 112–133. doi: 10.1177/1558689806298224.

Kilbourne, J., Scott-Webber, L., & Kapitula, L. R. (2017). An activity-permissible classroom: Impacts of an evidence-based design solution on student engagement and movement in an elementary school classroom. Children, Youth and Environments, 27(1), 112–134. https://doi.org/10.7721/chilyoutenvi.27.1.0112

Lever, J., Krzywinski, M., & Altman, N. (2017). Points of significance: Principal component analysis. Nature Methods, 14, 641–642. https://doi.org/10.1038/nmeth.4346

Murtagh, F., & Legendre, P. (2014). Ward’s hierarchical clustering method: Clustering criterion and agglomerative algorithm. Journal of Classification, 31(3), 274–295. https://doi.org/10.1007/s00357-014-9161-z

Preiser, W. F. E., Harvey, Z. R., & Edward, T. W. (1988). Post-Occupancy Evaluations. Van Nostrand Reinhold. Scott-Webber, L. (2000, December 10–13). Characteristics of a conceptual framework addressing the

environment of instructional space holistically: design, users, pedagogy and technology. In H. Thomas (Ed.), American Vocational Education Research Association Conference Proceedings (p. 54). New Orleans, LA.

Scott-Webber, L. (2004). In-sync—Environmental behavior research and the design of learning spaces. MI: The Society for College and University Planning. Retrieved from http://books.google.com/books/about/In_Sync.html?id=snRPAAAAMAAJ

Scott-Webber, L., Konyndyk, R., French, R., & French, J. (2018). Significant results: Space makes a difference for student academic engagement levels. European Scientific Journal, 14(16), 61. https://doi.org/10.19044/esj.2018.v14n16p61

Scott-Webber, L., Konyndyk, R., French, R., Lembke, J., & Kinney, T. (2017). Spatial design makes a difference in student academic engagement levels: A pilot study for grades 9–12. European Scientific Journal, 13(16), 5. https://doi.org/10.19044/esj.2017.v13n16p5

jel.ccsenet.org Journal of Education and Learning Vol. 8, No. 5; 2019

56

Scott-Webber, L., Marini, M., & Abraham, J. (Spring, 2000). Higher education classrooms fail to meet needs of faculty and students. Journal of Interior Design, 26(1), 16–34. https://doi.org/10.1111/j.1939-1668.2000.tb00356.x

Song, M. K., Lin, F. C., & Ward, S. E. (2013). Composite variables: When and how. US National Library of Medicine National Institutes of Health, 62(1), 45–49. https://doi.org/10.1097/NNR.0b013e3182741948

Tavakol, M., & Dennick, R. (2011), Making sense of Cronbach’s alpha. International Journal of Medical Education, 2, 53–55. https://doi.org/10.5116/ijme.4dfb.8dfd

Trochim, W. M. K. (2006). Correlation. WEB Center for Social Research Methods. Retrieved from https://socialresearchmethods.net/kb/statcorr.php

Trochim, W. M. K. (2006a). Convergent & Discriminant Validity. WEB Center for Social Research Methods. Retrieved from https://socialresearchmethods.net/kb/convdisc.php

Zimring, C. M., & Reizenstein, J. E. (1980). Post-Occupancy evaluation: An overview. Environment and Behavior, 12(4), 429–450. https://doi.org/10.1177/0013916580124002

Notes Research sponsored by DLR Group’s K12 Education Studio. With thanks to all of our respondents. Copyrights Copyright for this article is retained by the author, with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).